nano-banana-2-lite:official is a lightweight image creation model in the Google Nano Banana series, suitable for generating images from text and adjusting subjects, backgrounds, and styles with reference images. Positioned for 1K image creation, it focuses on concept drafts, asset variations, and interactive editing. For workflows that need to establish composition and visual direction before producing refined final work, Lite is a practical starting point.
Submit requests to the public API at api.acedata.cloud using the documented parameters, then use the results in your application.
api.acedata.cloudnano-banana-2-lite:officialInput parameters and result formats vary by service. Use the public API for this model and follow its guide for generation, task retrieval and editing operations.
Lite is positioned for 1K image creation; aspect ratios, task processing, and result fields are used according to this platform's image endpoint.
Organize prompts with descriptions of the subject, environment, lighting, composition, and style, and Lite can be used to explore different visual directions for the same idea. It is well suited to validating visual relationships early in the design process, such as comparing centered versus side-positioned products or bright versus dark backgrounds, rather than placing the entire task burden on a refined final result from the start.
When editing, provide the original image and distinguish in the prompt between elements to retain and elements to change—for example, keep the cup's shape and color while replacing only the tabletop and background. Multi-image combinations allow assets such as subjects and scenes to participate in creation together; clearly stating the purpose of each reference image helps reduce element mixing.
Seven aspect ratios cover square assets, landscape covers, and portrait displays. When choosing a ratio, describe both the subject position and the direction of negative space—for example, place the person lower in a portrait image and leave space at the top for a title. This designs the composition during generation rather than mechanically cropping the same image into different layouts.
Provide a product reference image, describe scenes such as a wooden table, by a window, or a minimalist interior, and require the product outline and main colors to be retained. The deliverable is a set of scene candidate images for discussion, making it easier for the team to compare backgrounds, lighting, and placement. When details such as packaging text or port locations are involved, they should be checked item by item against the physical product.
Provide the event theme, brand colors, and target canvas size; first generate cover images or poster backgrounds with different compositions, then adjust the atmosphere and elements around the selected direction. Lite is better suited for producing visual concept candidates; formal titles, dates, and rule text can be added during the layout stage to avoid making image generation responsible for all text proofreading.
Use the same character image as a reference, separately describe changes to the background, clothing colors, or visual style, and create drafts for stickers, mood images, and content illustrations. Clearly specify that facial features and key identifiers should be retained in every round, and check whether the result has deviated from the original design; reference images can help maintain the character's appearance, but do not lock every detail in place.
If the task focuses on finding the right composition, color palette, and scene, and 1K images are sufficient for preview and discussion, Lite can be prioritized. It is in a different tier from nano-banana-2: Lite emphasizes lightweight creation and rapid validation, while nano-banana-2 is a more balanced option. Choose based on delivery requirements rather than treating the two names as the same model.
If 2K or 4K delivery, complex detail presentation, or more refined brand visuals are needed, consider nano-banana-pro. Do not substitute for an upgrade by setting a higher resolution value for Lite. A more practical workflow is to use Lite to validate the concept first, then provide the selected composition requirements and reference materials to a model suitable for final production.
Choose generate for text-based creation; choose edit to modify existing assets, provide reference images with image_urls, and separately explain what to retain and what to change.
For /nano-banana/images, specify model=nano-banana-2-lite:official, action, and prompt; start with aspect_ratio=1:1, resolution=1K, and count=1; do not select 2K/4K for Lite.
Retrieve the image from data[].image_url; for asynchronous processing, query with task_id or receive a callback. When continuing edits, submit the selected image again and narrow the scope of changes in each round.
Create a square knowledge-sharing thumbnail with a bright yellow desk lamp and a closed book in the center, a blue background, clear empty space reserved on the right, and no long blocks of text.
Keep 1K and a simple layout, and use a small number of options to judge recognizability before clicking; count should not be interpreted as a guarantee of multi-reference editing or multi-round consistency.
It is suitable for 1K visual drafts, product scene-change mockups, campaign asset exploration, and reference image editing. It is especially useful for workflows that compare multiple creative directions before deciding on a final solution. If the task mainly evaluates print dimensions or extremely fine textures, higher-resolution models should be considered first.
Lite is intended for 1K image creation and 2K or 4K should not be treated as available output tiers. When higher resolution is needed, choose a model that supports the corresponding size, such as nano-banana-pro; changing only the resolution value in the request cannot replace model capabilities.
Use the edit action, place the reference image in image_urls, and clearly state in the prompt which subject features must be preserved and what needs to change in the background. Avoid changing too many attributes at once; complete the scene change first, then check whether the subject's outline, color, and details have changed in unwanted ways.
You should also adjust the composition instructions accordingly. For landscape images, emphasize left and right whitespace and the expanded environment; for portrait images, emphasize the subject's vertical position and space at the top. Lite provides seven aspect ratios, so choose a frame that fits the display placement and then describe the element layout, which is more targeted than cropping after generation.
Send a POST request to /nano-banana/images, specifying the model, action, and prompt; add image_urls when editing. The response provides a task identifier and image_url in data, which can be used for display or download. For asynchronous processing, use async and callback_url to arrange result delivery.
Choose a development approach or application suited to this model and follow its setup guide.
↗Python SDK Integration Guide
SDK installation, authentication and examples↗TypeScript SDK Integration Guide
SDK installation, authentication and examples
↗
↗
↗Pricing
$0.0172· per generation
You pay$13,584for 142,800 credits · $0.0951 / credits46% OFF
You're billed in credits. Pick a top-up tier to see its per-credit rate — the larger the top-up, the cheaper each credit.
This is only a basic call example. See the full docs for more parameters and advanced usage.
View Docs